Robust manufacturing system design using petri nets and bayesian methods

摘要

Manufacturing system design decisions are costly and involve significantinvestment in terms of allocation of resources. These decisions are complex, due touncertainties related to uncontrollable factors such as processing times and partdemands. Designers often need to find a robust manufacturing system design that meetscertain objectives under these uncertainties. Failure to find a robust design can lead toexpensive consequences in terms of lost sales and high production costs. In order to finda robust design configuration, designers need accurate methods to model variousuncertainties and efficient ways to search for feasible configurations.The dissertation work uses a multi-objective Genetic Algorithm (GA) and Petri netbased modeling framework for a robust manufacturing system design. The Petri nets arecoupled with Bayesian Model Averaging (BMA) to capture uncertainties associated withuncontrollable factors. BMA provides a unified framework to capture model, parameterand stochastic uncertainties associated with representation of various manufacturingactivities. The BMA based approach overcomes limitations associated with uncertainty representation using classical methods presented in literature. Petri net based modeling isused to capture interactions among various subsystems, operation precedence and toidentify bottleneck or conflicting situations. When coupled with Bayesian methods, Petrinets provide accurate assessment of manufacturing system dynamics and performance inpresence of uncertainties. A multi-objective Genetic Algorithm (GA) is used to searchmanufacturing system designs, allowing designers to consider multiple objectives. Thedissertation work provides algorithms for integrating Bayesian methods with Petri nets.Two manufacturing system design examples are presented to demonstrate the proposedapproach. The results obtained using Bayesian methods are compared with classicalmethods and the effect of choosing different types of priors is evaluated.In summary, the dissertation provides a new, integrated Petri net based modelingframework coupled with BMA based approach for modeling and performance analysisof manufacturing system designs. The dissertation work allows designers to obtainaccurate performance estimates of design configurations by considering model,parameter and stochastic uncertainties associated with representation of uncontrollablefactors. Multi-objective GA coupled with Petri nets provide a flexible and time savingapproach for searching and evaluating alternative manufacturing system designs.